Ground Resolution of the Imagery

Figure 6. The wind window: grey arrow indicating the wind direction and white arrows all possible kite movement. area beneath that semi-circle. Of course these flight characteristics may or may not be problematic and have to be considered on a site to site basis. During fieldwork, relatively long steering lines of 60 meter were used which enabled occupying less operating points and better overall coverage of the site. The disadvantage of these longer steering lines is the increase in drag on these lines, pulling the lines with the wind. Drag is caused by the wind blowing on the surface of the lines and although special lines were used, the sheer length of the used lines still created a significant amount of surface for the wind to create a drag. One of the more obvious signs of the drag is the curve the steering lines show between the operator and the kite. Whereas shorter steering lines without drag will show a much more straight line. The actual disadvantage of this drag is that the steering lines follow a longer path between kite and operator, and thus enable the kite to fly lower than theoretically possible. But the biggest drawback of the drag is the dampened and delayed response of the kite to the actions of the operator, leading to somewhat less control over the kite. 4.2 Operation of the camera In order to take pictures from the kite, one has to remotely operate it or set it to take pictures at a predetermined interval. Although the latter is the easiest and cheapest we found it has some serious disadvantages. The main disadvantage is that the operator can’t see which part of the site is photographed exactly. Judging the position of the kite from the ground, the operator only has an approximate idea of the area the camera captures. Also, the exact instance the camera takes a photograph isn’t known to the operator, only by approximation. To be sure that enough imagery of good quality was taken in the right direction more pictures were taken than was absolutely necessary. This meant that more flying hours were necessary. Another disadvantage of the used operation mode of the camera was that only the default automatic mode was available when shooting in interval mode. This meant that image sensitivity, aperture and exposure time could not be set by the operator but where chosen by the camera computer anew for every picture. This of course produced imagery with slightly varying characteristics regarding to aperture and exposure time which the camera computer thought was best for the imagery at the time. In general, though, the variation in the imagery characteristics turned out to be hardly problematic. Only some motion-blur in a few images could have been prevented by setting a better exposure time. The pictures with the motion-blur were excluded from further processing and again more imagery had to be taken with the kite to overcome the problem. 4.3 Flight Plan The flight plan was designed to fly at the same altitude as much as possible. Thus describing an arc, from left to right in the operators’ perspective with the kite. Due to some variety in the wind conditions during flight, the kite had to be flown higher or lower as wind conditions were sometimes more favourable at those altitudes. This brought some variety in the ground resolution of the imagery, but, as unstable winds could be largely omitted, the camera could to some degree be held free of unwanted motion in those stable wind conditions. Because of the length of the steering lines, only a few operating positions needed to be occupied in order to cover the whole area. As three quarries were recorded, only one operating point per quarry turned out to be necessary. This enabled the operator to completely focus on the kite and its flight plan without having to be aware of his step whilst flying the kite.

5. HIGH RESOLUTION DEM

5.1 Ground Resolution of the Imagery

Operated with two to three people the simple and low-cost setup as described above can acquire good aerial imagery Figure 7 with a high ground resolution to be further processed to a DEM or other surface model. Because the exact height of the camera is unknown, the exact ground resolution of the imagery cannot be determined precisely. On the basis of the length of the steering lines the theoretical maximum flying altitude of the kite was 60 meters above the ground. In this case the imagery was taken with a 28mm focal length 35mm frame equivalent which implies that the area This contribution has been peer-reviewed. doi:10.5194isprsarchives-XL-5-W4-77-2015 81 Figure 7. Typical aerial photograph of one part of the quarries. contained within the image was 0.035m x 60m 0.028m = 75 meters wide. With an image resolution of 4000 x 3000 pixels, the minimum ground resolution of one pixel can be estimated to be 75m 4000 = 0.019 meters. Most images however will not have been taken at that altitude but somewhere between 30 to 50 meters. At that altitude the ground resolution will have almost doubled to approximately 1 cm. per pixel. 5.2 Ground Control Points and Accuracy With the estimated ground resolution of the imagery between 1 and 2 cm. the accuracy of the ground control points gcp’s optimally will have to be half of that. The topometric instrument available to us at the time of fieldwork was a total station of the type Topcon GTS-601C, with kind permission of the Centre for Nabataean Archaeology at the Al-Hussein bin Talal Univ ersity of Ma’an. Through a local calibration of the instrument 1 it was found that the used total station setup was able to deliver an accuracy of 1 to 2 cm. depending mainly on wind conditions. Ideally, the uncertainty of the measurements should be in the range of 0.5 to 1 cm. but this would require either a different setup of the topometric measurements or extensive post- processing of the measurements. Since the main focus of this research was the suitability of the platform and establishing a workflow according to it, it was decided that the current accuracy of the measurements would suffice for the time 1 A calibration following the ISO 17123-5:2005 procedures for Electronic Tacheometers, executed by R. Emaus. being. Improving the quality of the gcp’s in a later stage of the research.

5.3 Image Processing